Monday, January 15, 2018

80 meter antenna

We built several Cricket 80a kits, which operate in the 80 meter amateur band.  Since we don't have any antennas specifically for this band, I decided to make one.  The design is a basic dipole fed with coax and a current balun at the feedpoint.

Here are the feedpoint parts: 

The coil is bifilar wound with 13 turns of 22 AWG speaker wire through a toroid.  It measures around 350 microhenries.  The extra loop is for a support cable to lift the feedpoint.  The enclosure is a watertight plastic container that has holes drilled for the SO-239 connector and mounting hardware.  I found that it was easier to control the drilling by hand (rather than by power drill), and a step drill made the holes cleanly deburred.

The two wires attach to the coax center and shield, while the other leads attach to the antenna.  It made sense to do the electrical work first...

... and then install it in the container.  I applied plastic epoxy to each of the pass-throughs before installing hardware in an attempt to seal out any water.  After the hardware was installed, I screwed in the antenna connections tight.

Here is the final enclosure, ready for the radiating wires.

Connecting from the hooks to the radiating wire segments is done with a short feeder segment of stranded wire.  Loops are soldered in the stranded wire while it's installed on the enclosure.  To avoid melting the plastic, I gripped a hemostat onto the plastic side of the junction to draw the heat.

Here are the feeder segments ready for the radiating wire segments. 

The next step was to design the radiating wire segments.

For the radiating structure, we are constrained by the feedline length (25 feet of RG-58) and the confines of our lot.  We're lucky that the antenna basically runs the length of one side of our property, basically touching the ground on one end due to ground slope.  Here is what NEC seems to suggest:
  • Height of antenna above ground: 2 meters
  • Length of each leg: 20 meters = 65 feet
Here is the impedance and VSWR according to NEC:

Here is the radiation pattern according to NEC, which clearly indicates that the main beam is vertical, which should be good for near vertical incidence skywave (superimposed on the antenna structure):

Given this plan, I measured out two runs of 67 feet each; better to cut long and trim than the other way around.

The wires are attached to the feeders.

Ready for installation!

Then I strung the works into position.  This took a while; after trimming off the excess wire, I got a near perfect SWR around 3.560 - 3.580 MHz, right where I wanted it.  As NEC predicted, the antenna seems to degrade higher in the band, with a 2.5:1 SWR around 3.800 MHz.

Thursday, January 4, 2018

Extreme IC repair

For Christmas, our family of hams got Cricket 80A transceiver kits.  They are not terribly hard to assemble and are good soldering practice. 

One problem we had with two of them was that the local oscillators did not start up.  It turned out that the 2N7000 MOSFETs are susceptible to electrostatic discharge (ESD) damage.  For whatever reason, the Q1 local oscillator transistor seems more vulnerable to this.

But in one of the kits, there was a problem with the audio amplifier IC, an NJM-2113D.  In the process of soldering and apparently removing, three pins broke off.  These are not exactly standard, at least they're not in my junk box. So although I put in an order for a replacement, Donna pointed out that I could probably fix it anyway.

The issue is that pin 1 (GND) broke off at the case, so I set about using a four fluted 1/8" endmill to cut the case to expose more of the pin.  I gripped the endmill in a collet in the headstock of my lathe.

My biggest concern was part-holding.  Fortunately, I was able to grip the IC in a toolpost, which also didn't break it.

After a few passes, I successfully exposed what seemed like enough of the pin to take solder.

I started by soldering the intact pins to the board and then the third pin, which was broken off, but not all the way at the case.  For this, I ran a piece of tinned copper wire through the hole, soldered it to the board, and then to the IC.

Then I ran another piece of tinned copper wire through the hole for pin 1 until it stopped on the exposed part of the IC, and soldered it to the board.

Finally, I soldered the new pin 1 to the IC.  I used a bit too much solder, but it's a good connection.

And, it works!  Here is the happy owner of the completed radio.

Sunday, December 31, 2017

My "usual" pants hem

I almost always wear "formal" pants, since I don't find jeans comfortable.  Since I'm short, I usually have to hem my pants.  While a simple "straight" hem is functional, it is not quite as sturdy as I need for every day usage.  The visible seam isn't quite as nice, also.

Since I had to re-hem a suit some years ago, I took apart a more fancy hem on the pants and developed my own technique.  Ironing is crucial.  I set the iron on its maximum temperature, since I'm setting creases rather than merely taking out wrinkles.  Depending on the fabric, this requires some caution so that the fabric isn't damaged.  Since I prefer wool over other materials, this is rarely a problem, but wool-synthetic blends require a more light touch with the iron to set a crease.

Here is now what I usually do... 

My inseam is 25", so everything is referenced off of that.  I start by cutting the pants to somewhere between 28 1/2" and 29" (inseam plus 4"). 

Then I fold this in to 28" (inseam plus 3") and iron.  (I don't turn the pants inside-out, so the picture below is of the outside of the pants along the inseam.  The extra fabric you see from 28" to 28 1/4" is the inside of the other side of the same pant leg.)

I machine stitch this edge.  Thread color does not matter, since it will be hidden.

I bury the bitter ends of the thread on the inside of the seam, so they can't catch on anything.

Then I iron another crease, now at 26" (inseam plus 1").

And before any further stitching, I fold and iron at 25" (the final inseam).  This almost always requires several passes of the iron since there are many layers of fabric to crease.

At this point, the hem needs to be anchored with a stitch.  But since I don't want the thread to show through the outer layer, I don't stitch this with the machine.  So I hand stitch through all layers but the outer layer.

Here is the view from the inside of the pants...

... you can see no thread on the outside ...

... the stitching holds together here.

Here is the finished product.

Thursday, December 28, 2017

Transistor oscillator

I have been looking for a simple, easy to understand oscillator circuit that uses a single transistor.  Well, I found one!  Twice, it seems.  The first place was in a classic vacuum tube circuit, as described in

Morecroft, Elements of Radio Communication, Wiley, 1929.

After suitable edits to make it transistorized, I then found the same circuit in a hand-drawn schematic that my father had squirreled away in the 1944 edition of the ARRL Handbook. Certainly not new!

For my and Edwin's benefit, I built the circuit using snap circuits.

Here's the schematic:

The different parts of the circuit are indeed easy to understand:
  • The tank resonator is an inductor-capacitor (LC) circuit, which sets the frequency.
  • The keying turns on and off the oscillator.
  • The bias ensures that the transistor is turned "on" and not saturated.
  • The transformer affords feedback from the tank to the input of the transistor.  It also contains the inductor part of the tank.  In the snap circuits version, we don't have much control over this (it's in a plastic package), but too little inductance will cause the circuit to fail to oscillate.
  • The 200 ohm feedback resistor sets the gain of the transistor amplifier.  Reducing the resistance increases the gain, which drives it harder.  Increasing the resistance makes for a cleaner signal, but can also stop the oscillations. It seems to stop around 1000 ohms or so, but with the 200 ohms it is noticeably overdriven.
  • The signal purity can be increased by selecting a higher voltage for the power supply.  I got it to work with 3 volts, but the output was more clipped.

Clock 3 cased and installed

Clock 3 is finally complete, and is now installed in an oak frame/case in my office.  The case has a matching French cleat so the movement is easy to remove for debugging.  I also installed a dial that is perched on pins on the shelf, which allows it to be easily removed by lifting it off the movement. 

It is now driven by a 10 pound bag of lead pellets (intended for scuba diving) with a two-fall pulley.  It runs for 16 hours on a wind.  This short run time is well enough so that it shouldn't bother my office neighbors.  It's a bit noisy, but now that it is no longer mounted on a hollow cavity, it's quieter than before.

Moving the clock from my cool, damp basement to the warm, dry office did require some adjustments...  The movement's frame has the grain going horizontally, which meant that it shrank vertically a bit.  This caused the impulse pin to bind on two things:
1. The detent tip, which required shifting the detent back slightly
2. The trailing edge of one escape wheel tooth, which required a small amount of filing.

The great wheel also fouled on the frame -- an indication that my "fancy" milled frame wasn't a good idea -- since it appears that the frame has shrunk vertically.  To compensate, I carved the milling back further.

The great wheel pivot, which I had previously needed to move (and I wondered why!) had to be set back in its original position as well.

Additionally, many friction-fit parts loosened and required the use of super glue to anchor:
1. The impulse pin
2. The hour finger
3. The dial pins 

Loop antenna

Since HF antennas tend to be rather large, I wanted to try a small resonant loop.  After doing some research, I found a few sites that seemed rather detailed.  This one seems about the best.

I constructed mine from 3/8" copper tubing, and it is mounted on a camera tripod base.

It is fed using a gamma match.

The tuning is accomplished by a scavenged broadcast FM tuning capacitor and a homebrew air-variable butterfly capacitor.

The antenna can be tuned on the 20 meter through 10 meter bands.  Here I am using it on 20 meters...

Wooden hygrometer

Based on the issues of wood expansion with my clocks, a natural project would seem to be a wooden hygrometer.  At least, it ought to respond about as fast as the wood in the clocks responds, which appears to be around a few days. 

The design is fairly simple, with two oak sticks of roughly the same length being placed parallel to each other and rigidly anchored at one end.  One of the strips has the grain running lengthwise, while the other has the grain running perpendicular to that.  The two free ends are pinned to a needle.  I placed the parallel-grain strip on the bottom, so that the needle moves to the right as the humidity increases.  With this particular item, fully soaking it for a week in water runs the needle all the way to the right of the scale.

But the typical indoor humidity is closer to the first picture.  The needle usually moves around 1/4"-1/2" with daily variations.